Discovery of High-Performance Extremophiles and Extremozymes Using Machine Learning and Structure-Based Clustering

Area of Science:

• Microbiology
• Bioinformatics
• Machine Learning

Background:

• Extremophiles are vital for biotechnology and understanding life's limits.
• Traditional extremophile identification methods are inefficient and labor-intensive.

Purpose of the Study:

• To develop an efficient machine learning model for predicting extremophile characteristics.
• To identify novel extremophilic species and extremozymes.

Main Methods:

• Developed iExtreme, a Support Vector Machine (SVM) model using k-mer and codon features from genome sequences.
• Trained the model on 1030 extremophilic genomes.
• Employed structure-based protein clustering for extremozyme discovery.

Main Results:

• iExtreme achieved high accuracy in identifying halophiles (0.988), thermophiles (0.939), and pH-philes (0.938).
• Discovered 520 novel extremophilic species and 5255 genomes.
• Identified novel extremozymes, including d-psicose 3-epimerases and α-amylases.

Conclusions:

• iExtreme is a powerful and accurate tool for extremophile identification.
• The model facilitates the discovery of novel extremophiles and enzymes.
• This work advances extremophile research and biotechnological applications.